Configure Google Cloud Build using Docker

You can install packages using your console. For example, in macOS:

terminal
brew install tesseract

In Google Cloud Functions, you can't install system packages via terminal. Thus, if you want to have a specific system configuration, you'll need to tell somehow Google Cloud how to configure your environment.

That's why dockerizing is useful, because it allows you to set a system with a concrete configuration (container). Google Cloud Build allows you to use a Docker file to configure your application.

Let's assume that you want to deploy the following Python app:

python
from flask import Flask

app = Flask(__name__)

@app.route('/')
def hello():
    return 'Hello world'

if __name__ == "__main__":
    app.run(debug = True, host = '0.0.0.0', port = int(os.environ.get('PORT', 8080)))

The app will return 'Hello world' when you access to the route / from your browser.

In addition to the app, create a folder with the following files:

terminal
├── hello_world
|   ├── app.py
|   ├── requirements.txt
|   ├── Dockerfile
|   ├── cloudbuild.yaml

You can specify the Python dependencies to be installed by Google Cloud via requirements.txt . The following dependencies won't be used, but we are going to install them anyway for this example:

requirements
layoutparser
layoutparser[ocr]

You can specify the desired container's state via Dockerfile:

docker
FROM python:3.7
RUN mkdir /app
ADD . /app
WORKDIR /app
RUN pip install Flask gunicorn
RUN pip install -r requirements.txt
RUN apt-get update && apt-get install tesseract-ocr
CMD exec gunicorn --bind :$PORT --workers 1 --threads 8 app:app

Notice how Dockerfile will configure the system with Python 3.7, Flask, gunicorn, layoutparser and layoutparser[ocr]. Moreover, it'll install tesseract-ocr, not as a Python dependency, but as a global package.

At this point, you can configure the cloudbuild.yaml file:

yaml
steps:
- name: 'docker'
  args: [ 'build', '-t', 'gcr.io/$PROJECT_NAME/helloooooo', '.' ]

images:
- 'gcr.io/$PROJECT_NAME/helloooooo'

tags: ['gcp-cloud-build-sample-build']

To build the container:

terminal
gcloud build submit

At this point, you can use Cloud Run to deploy the container:

terminal
gcloud run deploy --image gcr.io/$PROJECT_NAME/helloooooo

After deploying the container, you'll receive an URL to access the service.